| This paper studyed on the Tianluoshan tunnel of the Zengcong Highway in Guangzhou,based on the scene measured data,established a method about the tunnel back-analysis of displacement and surrounding rock stability analysis models, and applied in practical engineering, This paper studies the content and results are:1.Basis on collecting the Tianluoshan tunnel detailed information,formulate reasonable control monitoring program,focused on regression analysis and statistical analysis of the dome subsidence and peripheral convergence in the monitoring data, analysis the relationship of the surrounding rock deformation and the space and the time,forecast the surrounding rock deformation trend,provide a reasonable support time for second- lining;2. According to BP neural network and the basic principles of genetic algorithms, optimized the weights and thresholds of BP neural network with genetic algorithm, established a GA-BP neural network model for back-analysis of tunnel surrounding rock, which based on MATLAB 7.0 and GAOT toolbox; Solved the forward analysis of the tunnel surrounding rock with FLAC3D, established the parameters of training samples with the method of orthogonal design, calculated the parameters of the tunnel rock based on GA-BP neural network;3. According to the basic characteristics of FLAC3D software and mechanical principles of tunnel construction process, established a simulation model of three-dimensional tunnel excavation based on FLAC3D; Filled the results of parameters into the numerical model,calculated by the surrounding rock and the tunnel lining in all stages of force change,from the mechanical point of view of the surrounding rock and the tunnel support structure in the process of excavation stability issues, and comparated it with measuring and monitoring results,which provided a foundation for the follow-up tunnel design, construction and monitoring. |
PDF Full Text Request